AI may trigger catastrophic loss of control very soon: OpenAI's new safety hire

AI safety pioneer Paul Christiano joined OpenAI nonprofit board, serving on Safety and Security Committee to support safety oversight. "Rapid acceleration in AI capabilities leads to catastrophic and irreversible loss of control in the very near term," Christiano said. Earlier, Anthropic researcher Evan Hubinger claimed there's an over 10% chance that AI could kill all humans within next decade.